We can pretty easily aggregate stats from mediawiki history and pageviews for wikis inside incubator.wikimedia.org. This way it would be more practical to follow the progress of the 300+ wikis that are currently living there, and better determine wikis in danger and wikis ready to graduate.
A little oozie job generating a dataset with all metrics would be a good starting point.
Thanks for creating this, @fdans!
The tool we currently use to monitor activity in the Incubator is "Catanalysis" by @Pathoschild. For the Sakizaya Wikipedia (recently approved), it looks like this.
It works quite all right for getting an overview for a single wiki, but as far as I know there's no way to get "the whole picture" to see which wikis are most active, so we basically have to monitor recent changes and see which wikis have activity to get an idea of which projects could be getting closer to approvable, which makes it easy to accidentally overlook some. (But please do correct me if I'm wrong @MF-Warburg and @StevenJ81.)
The most important metrics we look at are: (1) long-term activity (as in month-over-month), and (2) how many contributors are involved (month-over-month). This would then need to be sorted by test wiki prefix (<project code>/<language code>/)
That's pretty much it, @jhsoby. I don't promise that when we see an alpha version of this, some further refinements won't come to us. But this is a good place to start. (Calling also @Ooswesthoesbes .)
Yes, that's the basics. The advantage of the catanalysis is the number of bytes added, as some contributors make a lot of low change edits, while others make a few bigger edits. But if this tools main function will be to see which wikis are active, and which ones not, it is not necessary.